What all can you find out from DNA anyway?

Have you all heard of "23andme" where you can get your DNA tested from a saliva sample for all kinds of things? A relative of mine had it done to see his ancestry among other things. No real surprises: predominantly European but with an unexpected tiny smidgen of South Asian. (Who knows how far back that goes?) The European was broken down into many subgroups, mostly all Western European nationalities, but also including Ashkenazi Jewish. (I wonder why no Sephardic Jewish because that should be in there somewhere.)I'm surprised at the detail. I had no idea you could get that much specific information about ancestry from DNA. Is it possible to explain how this is possible?DNA Genetic Testing For Health, Ancestry And More - 23andMe

That's a good description of how the ancestry of ancient books is reconstructed, but I don't see how "errors" get into the DNA question. I'm just amazed that there would be anything IN the DNA that could actually indicate, say, a 2% South Asian ancestry. Which genes or other part of the genome could possibly tell us that? Doesn't that amaze you too?

How about sparing us the ad hominem and saying something substantive?

Welcome to the Abnormal Club, Chapter EvC.

Yes I'm slow but it was dawning on me that the idea is that some mutations mark particular groups of people and that's how they can give us a genetic portrait of which groups we're related to. Although Tangle seems to believe I'd be offended at the idea of mutations, that's a bad misreading of the contexts in which I've objected. I object to the idea that they are normal events that make normal alleles, I object to the idea that they contribute anything genuine to genetic diversity and so on. But there's no reason I should object to their identification of people groups; in fact that's an unexpected benefit from an error.So are you all saying that it's mutations and mutations only that give us the ability to trace our ancestry?Edited by Faith, : No reason given.

Well, without mutations it wouldn't all be the same, we'd each have our own unique genetic combinations.But I know that's a nitpick because tracing such combinations to ancestors would probably be practically impossible, and unique mutations do make a neat way to pinpoint groups.

Of course I wonder how reliable the test is. What if there are people groups that have no identifying mutations?Or, here's a specific question: the person who was tested has a Jewish ancestor who is half Northern/Eastern European/Ashkenazi and half Sephardic (Portuguese in this case) but the percentage given makes the ancestor about 97% Ashkenazi. Should this be explained as there being no particular Sephardic mutation? So that it's all just Jewish DNA? Or could it mean that the Sephardic ancestors had a lot of Ashkenazi ancestry they didn't know about? (A little odd in this case because they have genealogies that go back to the fifteenth century with Sephardic names). And should that unaccounted for 3% be supposed to come from a nonJewish ancestor? Perhaps even that small South Asian percentage?Also said person has a grandmother who is French. But the percentage of French in his test is minuscule, like 1.2%. Does that mean the French grandmother doesn't know her real genetic background or what? Edited by Faith, : No reason given.Edited by Faith, : No reason given.